Jaramillo Fernando Mosquera, Oliveira Tiago Marcelo, Silva Pedro Enrique Ayres, Trindade Pedro Henrique Esteves, Baccarin Raquel Yvonne Arantes
Department of Internal Medicine, School of Veterinary Medicine and Animal Science, University of São Paulo, São Paulo, Brazil.
Equine Centre, Melbourne Veterinary School, Faculty of Science, University of Melbourne, Werribee, VIC, Australia.
Front Vet Sci. 2023 Aug 10;10:1189846. doi: 10.3389/fvets.2023.1189846. eCollection 2023.
Horse racing is a major sport practiced worldwide. The environment to which horses are exposed during race meetings can influence their behavior. However, to the best of our knowledge, a method for assessing a horse's response to its surroundings during the pre- and post-race periods has not yet been reported. This study aimed to create a standard list of descriptors for use in a qualitative behavioral assessment (QBA) focused on assessing the emotional expressivity of horses before and after racing events.
Seventy pre- or post-race 30-second videos of horses were randomly selected from our database of 700 videos. A panel of 8 experienced equine sports medicine specialist veterinarians watched a 60 min presentation on QBA. The panel then watched all videos randomly, simultaneously, individually, continuously, and without any verbal interaction, describing the descriptors related to the emotional expressivity of the horse after each video using a method known as free-choice profiling (FCP).
The initial selection of descriptors was based on those indicated by more than one evaluator in the same video, or descriptors with more than 20 occurrences. The second selection was performed based on the content validity index (CVR) to select the descriptors retained in the previous step. Another panel of six veterinarians scored each of the descriptors retained for content validity on a visual scale. Interobserver reliability was estimated using the intraclass correlation coefficient (ICC) and its respective 95% confidence intervals (CI). A natural language processing (NLP) algorithm was used to analyze the behavior (positive or negative polarity) of the descriptors based on the lexicoPT package of R software.
DISCUSSION/CONCLUSION: NLP analysis considered the descriptors "agitated," "troubled," "restless" and "irritated" to have a negative polarity, while "focused," "relaxed" and "peaceful" had a positive polarity. In the principal component analysis (PCA), descriptors in a negative state were associated with each other and inversely associated with descriptors in a positive state. We conclude with a fixed list of descriptors to be used in a QBA to assess emotional and welfare expressivity in racehorses' pre- and post-race environments.
赛马是一项在全球范围内开展的主要运动。马匹在赛事期间所处的环境会影响它们的行为。然而，据我们所知，尚未有报道称有一种方法可用于评估马匹在赛前和赛后对周围环境的反应。本研究旨在创建一份标准描述符列表，用于定性行为评估（QBA），重点评估赛马在赛事前后的情绪表达。
从我们700个视频的数据库中随机选取70个赛前或赛后时长30秒的马匹视频。由8名经验丰富的马运动医学专家兽医组成的小组观看了一场关于QBA的60分钟演示。然后，该小组随机、同时、单独、连续且无任何言语互动地观看所有视频，使用一种称为自由选择剖析（FCP）的方法，在每个视频播放后描述与马匹情绪表达相关的描述符。
描述符的初步筛选基于同一视频中被不止一名评估者指出的描述符，或出现次数超过20次的描述符。第二次筛选是根据内容效度指数（CVR）进行的，以选择上一步保留的描述符。另一个由六名兽医组成的小组对保留的每个描述符的内容效度进行视觉评分。使用组内相关系数（ICC）及其各自的95%置信区间（CI）估计观察者间信度。基于R软件的lexicoPT包，使用自然语言处理（NLP）算法分析描述符的行为（积极或消极极性）。
讨论/结论：NLP分析认为描述符“激动的”“不安的”“烦躁的”和“恼怒的”具有消极极性，而“专注的”“放松的”和“平静的”具有积极极性。在主成分分析（PCA）中，处于消极状态的描述符相互关联，与处于积极状态的描述符呈负相关。我们得出了一份固定的描述符列表，用于QBA中评估赛马赛前和赛后环境中的情绪及福利表达。
